Background. To assess the role of flourodeoxyglucose-positron-emission tomography (FDG-PET) scan in staging patients with nonsmall cell lung cancer (NSCLC). Methods. We prospectively studied 400 patients with NSCLC. Each patient underwent a computed tomography (CT) scan of the chest and upper abdomen, other conventional staging studies and had a FDG-PET scan within 1 month before surgery. All suspicious N2 lymph nodes by either chest CT or by FDG-PET scan were biopsied. Patients that were N2 and M1 negative underwent pulmonary resection and complete thoracic lymphadenectomy. Results. The FDG-PET had a higher sensitivity (71% vs 43%, p < 0.001), positive predictive value (44% vs 31%, p < 0.001), negative predictive value (91% vs 84%, p = 0.006), and accuracy (76% vs 68%, p = 0.037) than CT scan for N2 lymph nodes. Similarly, FDG-PET had a higher sensitivity (67% vs 41%, p < 0.001), but lower specificity (78% vs 88%, p = 0.009) than CT scan for N1 lymph nodes. FDG-PET led to unnecessary mediastinoscopy in 38 patients. FDG-PET was most commonly falsely negative in the subcarinal (#7) station and the aortopulmonary window lymph node (#5, #6) stations. It accurately upstaged 28 patients (7%) with unsuspected metastasis and it accurately downstaged 23 patients (6%). Conclusions. The FDG-PET scan allows for improved patient selection. It more accurately stages the mediastinum, however there are many false positives lymph nodes and it may be more likely to miss N2 disease in the #5, #6, and #7 stations. A positive FDG-PET scan means a tissue biopsy is indicated in that location. © 2003 by The Society of Thoracic Surgeons.